Abstract
The national government instated a mandatory blending policy to facilitate the uptake and estab-lishment of a biofuels sector in South Africa. Uncertainty exists, however, regarding the implica-tions and effects of producing biofuels within the Western Cape province, as part of a strategy of the province to transition to a green economy. This investigation was carried out as an effort to simulate the biofuel production within the Western Cape under certain project and policy considerations. A system dynamics model was developed to identify key strategic intervention points that could strength-en the business case of biofuel production. The model showed a feasible business case for bioethanol production, with the best case showing an internal rate of return of 23% (without govern-ment subsidy), and an emissions reduction of 63% when compared with coal. It is recommended that special consideration be given to the location of bioethanol production facilities, as operational costs can be minimised by incorporating invasive alien
land-clearing schemes as part of the bioethanol pro-duction. The model further showed that medium-to-large-scale biodiesel production in the province is not feasible under the given model assumptions, as the positive effects of local biodiesel production do not justify the required government subsidy of ZAR 4.30 per litre. It is recommended that a different approach be investigated, where multiple on-site small-scale biodiesel production facilities are used, thus utilising multiple feedstock options and min-imising capital expenditure.
Keywords: green economy; transportation; blend-ing; uncertainty; complexity
* Corresponding author: Tel: +27 (0)21 808 9530; Email: acb@sun.ac.za
Journal of Energy in Southern Africa 28(1): 1–12 DOI: http://dx.doi.org/10.17159/2413-3051/2017/v28i1a1457
Published by the Energy Research Centre, University of Cape Town ISSN: 2413-3051 http://journals.assaf.org.za/jesa
Sponsored by the Department of Science and Technology
Implications of biofuel production in the Western
Cape province, South Africa: A system dynamics
modelling approach
Willem Jonker
1, Alan Colin Brent
1,2*, Josephine Kaviti Musango
3,
Imke de Kock
11.Department of Industrial Engineering, and the Centre for Renewable and Sustainable Energy Studies, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa.
2.Sustainable Energy Systems, School of Engineering and Computer Science, Victoria University of Wellington, Kelburn, Wellington 6012, New Zealand.
3.School of Public Leadership, and the Centre for Renewable and Sustainable Energy Studies, Stellenbosch University, Private Bag X1, Matieland 7602, South Africa.
1. Introduction
Sustainable development is a well-established con-cept, and recognises the interdependence between the environment, the economy, and human society. A variety of concepts and policies were established to facilitate the incorporation of sustainable devel-opment, where one of the more recent concepts is that of a green economy. The United Nations Environment Programme (UNEP) describes the green economy as: ‘an economy that will result in improved human well-being and social equity, while significantly reducing environmental risks and ecological scarcities’ (UNEP, 2010). The green economy principle is especially applicable to devel-oping countries, like South Africa, where manage-ment and use of natural resources play a critical role to sustain economic growth and the population’s livelihood.
The South African government, as part of a sus-tainable development initiative, described its vision to transform the South African economy to a low-carbon, environmentally sustainable, and resilient economy in the National Development Plan (NDP) (National Planning Commission, 2012). The NDP stipulates various policies concerning sustainable transition, at a national level, which need to be implemented at a provincial level. The Western Cape Government, subsequently, published a framework (‘Green is smart’) to guide the province’s change into ‘the leading green economic hub of the African continent’ (Western Cape Government, 2013). The framework identifies clean energy and transport efficiency as two of the major role players in establishing a green economy. The need for clean energy, coupled with the nation-al mandate to diversify the energy supply mix, established biofuel as an important role player in the transition a green economy. The mandatory blending policy (officially effective October 2015, but not yet realised) calls for fossil fuel-based petrol and diesel to be blended with 2–10% bioethanol and 5% biodiesel, respectively (Department of Energy (DOE), 2012).
On a national level, manufacturers were granted biofuel production licences. The Western Cape, however, did not see any notable interest from investors to establish a large-scale biofuel produc-tion facility. The lack of interest may be due to pric-ing policy structure uncertainties, as well as general uncertainties regarding the feasibility and effects of large-scale biofuel production in the Western Cape (DOE, 2014). In order to address the uncertainties involved, biofuel production may be recognised as a complex system with various interactions, interre-lationships and interdependencies of which the effects are often unforeseeable and unintended, and identifying the most applicable approach to analysing large-scale biofuel production in the Western Cape, as a complex system, is crucial.
The objectives of this investigation, within the context of the Western Cape, were then to:
1. identify drivers, constraints and opportunities of commercial biofuel production;
2. investigate methods to inform strategic decision-making and use the most appropriate method to identify and advise on key strategic areas with the aim of strengthening the business case of biofuel production;
3. provide recommendations on the way forward for the province to form part of the mandatory blending policy; and
4. conclude on the usefulness of the technique identified to assess biofuel production as part of a green economy transition.
2. Literature review
A review of literature concerned with biofuels is pre-sented in order to understand and conceptualise the factors driving and influencing biofuel production. Biofuel production is subsequently identified as a complex system, and various techniques used to aid in capturing the complex nature of systems are reviewed; and the most appropriate technique is selected to simulate and evaluate the effects of bio-fuel production within the province.
2.1 Biofuels as a complex system
The need to diversify the energy supply mix in South Africa was first established in 1998 when the white paper on energy policy of South Africa (Department of Minerals and Energy, 1998, 2007) was released and became one of the focus areas in government’s vision to promote and implement renewable energy in South Africa. The large-scale utilisation of renewable energy sources could have numerous advantages, including: reduced green-house gas emissions, the promotion of small and micro enterprises to alleviate poverty and unem-ployment, and attracting private investment (both foreign and local) for the commercialisation of renewable energy production. The National Biofuel Industrial Strategy, released in 2007, mandates the blending of the national liquid fuel supply with 2% (per volume) of biofuels by October 2015. The 2% penetration level was revised from an original 4.5% due to concerns regarding the impact on food secu-rity, and the effect on energy density of blended fuels (Department of Minerals and Energy, 2007). Although an estimated 20% of the country’s fuel supply is refined in the Western Cape, the province currently has no approved large-scale commercial activities aiming to produce biofuels (DOE, 2014; South African Petroleum Industry Association, 2015). If the province were to form part of the blending mandate, it would have to start producing biofuel locally or purchase biofuel from one of the approved commercial plants elsewhere in the coun-try. Numerous studies have completed technology
and feedstock analyses for biofuel production in the Western Cape, where canola and triticale were identified as the most feasible feedstocks to produce biodiesel and bioethanol respectively (Nolte, 2007; Davis-Knight et al., 2008; Amigun et al., 2011).
Growth in the biofuels sector is driven primarily by support for renewable energy, due to advantages such as boosting the agricultural sector and limiting global warming. Extensive networks of policies, resources, regulations and drivers are, however, involved in the biofuels industry, and major con-cerns include: water limitations, food security, land value and availability, biofuel quality, crop selec-tion, fuel levies and subsidies, and the effects of large-scale production on the agricultural sector, including employment creation. The Western Cape government’s drive to transition to a green econo-my created a need for a holistic evaluation that con-siders the dynamic interaction of adverse effects of biofuel production, such as: job creation, land implications, emissions, and the bottom line costs involved to comply with the mandatory blending policy.
2.2 Systems thinking
Systems thinking is proposed by Loorbach (2010) as an approach to capture the complexities involved in persistent society problems, whereby possible solutions to sustainability challenges are likely to lead to further complexities. This means that sys-tems with a high level of interconnectedness and interdependence on various levels would not be accurately approached through linear, chronologi-cal thinking. Systems thinking, or complex systems theory, can be used as a language to describe the complex interactions and patterns between system components (Loorbach, 2010). Systems thinking is, therefore, an approach that considers all of the pos-sible influencing factors in a system and establishes their interconnectedness, mainly through mod-elling. Figure 1 indicates the phases of a systems thinking modelling approach, as described by Maani and Cavana (2007). Having identified biofu-el production as a complex system, due to the vari-ety of interactions on various levels, a systems approach is used in an attempt to accurately describe the various expected outcomes of biofuel production.
2.3 Dealing with complex systems
A variety of methods have been developed that can be used to analyse and represent complex systems. Most of these methods were developed to solve specific problems and progressively adapted to dif-ferent applications. Not all methods are, therefore, equally useful under similar conditions. In order to determine the most suitable method to represent biofuel production as a system with various dynam-ic relations and outcomes, the pitfalls and advan-tages of the most relevant methodologies were investigated. Bassi (2014) highlighted that data frameworks and modelling methodologies have to be considered to generate and analyse simulations of social, economic and environmental scenarios. Some of the dynamic modelling methodologies proposed by Bassi include econometrics, optimisa-tion, and system dynamics. Through surveying liter-ature from various authors that applied the respec-tive methodologies extensively, it was possible to assess each of the relevant methodologies to identi-fy which would be best suited to simulate biofuel production as part of a green economy given the following criteria:
• Problem identification: Does the methodology assist with problem conceptualisation within the confines of the system boundary?
• Flexibility: Can the model methodology incor-porate inputs from various platforms?
• Outcome accuracy: How reliable is the outcome of the simulation compared with the amount of computation time required?
• Identification of effect over time: Does the simu-lation deliver results as a function of time and can it model future projections?
Table 1 summarises the methodology assess-ment based on the work of various researchers.
System dynamics was ultimately selected as an approach to represent the complex real world prob-lems involved in biofuel production, due to its abil-ity to accurately represent interactions and intercon-nections from a variety of sectors and disciplines with relative ease. System dynamics uses a top-down approach and incorporates systems thinking to describe, model, simulate and analyse complex systems. System dynamics can adapt to parameters and inputs from various platforms and is transdisci-plinary-oriented, meaning that it is suitable for research efforts conducted by investigators from dif-ferent disciplines working jointly to create new con-ceptual, theoretical, methodological, and transla-tional innovations that integrate and move beyond discipline-specific approaches to address a common complex problem. It is also based on principles and techniques derived from control and feedback sys-tems, which gives it the distinct structure that direct-ly incorporates feedback, stock and flow structures (Pruyt, 2013).
Figure 1: Phases of systems thinking (Maani and Cavana, 2007).
Problem
structuring Causal loopmodelling modellingDynamic
Scenario planning &
modelling
3. Methodology
With the appropriate modelling technique identi-fied, a model had to be developed and used to ensure that the complexity of biofuel production as a system is accurately portrayed within the model. The framework of Maani and Cavana (2007), as shown in Figure 1, was favoured for its versatile and transparent nature, which uses a systematic process that aids in facilitating model validity and ensuring repeatability.
3.1 Problem structuring
The main drivers and areas of concern affecting the biofuels industry were identified through surveying the latest Western Cape specific literature and con-sultation with industry experts (Amigun et al., 2012; Nolte, 2007; Green Cape, 2015; Fore et al., 2011). The following key aspects were consistently men-tioned and directly influence the feasibility of biofu-el production and the required compliance with the mandatory blending policy:
• Biofuel production plant: The type and size of plant influences building cost, resource use and production efficiency, which directly drives the capital expenditure and operational expendi-ture.
• Agriculture: The type of agricultural crop used has an impact on production as one has to adju-dicate based on feedstock cost, availability and biofuel yield of the specific crop.
• Food security: The country’s food security is dependent on the local agriculture sector, and the allocation and selection of crops to produce
biofuel has to be done in a manner that will not negatively impact on food security.
• Land resources: A limited amount of land area in the Western Cape (19% of the province’s total land) is suitable for crop growth and it is impor-tant that crops grown for biofuel production do not encroach on land allocated for food crop production. The size of the plant and the type of crop used for production directly influences the land requirements.
• Water resources: Producing biofuel places addi-tional stress on water resources through increased agricultural activities, process water, and make-up steam and cooling water.
• Energy requirements: Biofuel (especially bio-ethanol) production is energy-intensive and, in the midst of South Africa’s energy crises, it is necessary to consider alternative energy sources and production strategies.
• By-products: Depending on the production pro-cess used, the by-product of biofuel production could strengthen the business case. Dried dis-tillers’ grains and solubles (DDGS) is a by-prod-uct of both bioethanol and biodiesel prodby-prod-uction processes and is commonly used as animal feed. Glycerine (used in the cosmetic and pharmaceu-tical industries) is also produced as a by-product in biodiesel production. The local market for glycerine is, however, saturated due to a large global over-supply, and glycerine then consti-tutes a waste management challenge.
• Biofuel sales price: The financial feasibility of biofuel production ultimately depends on
in-Table 1: Modelling methodology comparison (based on Epstein, 2014; Banos et al., 2011; Bassi, 2014; Pruyt, 2013; Brailsford et al., 2014; Helbing and Balietti, 2013; Snijders et al., 2010; Le Novère, 2015).
Methodology Main strength Main weakness Problem Flexibility Outcome Identification identification accuracy of effect over
time
Econometrics Based on historical trends Absence of feedback effects 3 3
Optimisation Gives an accurate estimation of Does not identify the drivers 3 whether a target can be reached contributing toward reaching
(given set constraints) targets
System Simulation driven by root cause Detailed input parameters, 3 3 3 3
dynamics and effect (accurately capturing or data, need to be obtained dynamics and feedback effects) across all sectors
Discrete Simulation (random or ordered) Rigid sequencing of events and 3
event event -driven systems, where stochastic nature delivers varying simulation entities have to take part solutions (time-consuming to
in processes run various simulations)
Agent-based Captures emergent phenomena Computation-intensive and 3 3 3
modelling (by working on agent level) limited capability to integrate actors from different sectors on
the same platform
Network Identifies the most important actors Limited ability to receive, 3 3 modelling in a seemingly chaotic complex manipulate and produce
system and shows connections quantitative values or between entities that would parameters otherwise not easily be identified
come generated from sales. The sales price is normally defined as a percentage (based on the ratio of biofuel to fossil fuel energy density) of the basic fuel price, which is determined by the crude oil price and rand-US dollar exchange rate.
In addition to the influencing factors in the bio-fuel sector, indicators are used to assess the impact and feasibility of different biofuel production sce-narios (Jonker, 2015). These indicators are identi-fied and described in Table 2.
Once the drivers and indicators were identified, they could be schematically represented, in order to find the points of relation from interconnecting sec-tors and to serve as a visual aid and expand on pos-sible exogenous drivers and endogenous effects. 3.2 Causal loop modelling
In system dynamics the visual representation of drivers and indicators is done in the form of a causal loop diagram (CLD). The CLDs consist of balancing (B) and reinforcing (R) loops, as indicat-ed in Figure 2, where a balancing loop shows two
Table 2: Model indicators.
Indicator Description Units
Environment
Emissions Change in air emissions due to biofuel sector kg CO2/ year
Land use Agricultural land required for biofuel feedstock cultivation Ha
Water use Water resources required for biofuel production litre / year
Electricity use Energy used to produce biofuel kW
Biofuel by-product Amount of DDGS produced per year litre / year
Social
Employment Additional employment creation due to biofuel production Person Economic
Investment into biofuel Capital investment needed to produce biofuel rand Operational cost The running cost of a production facility to produce biofuel rand / year Subsidy required The subsidy required per litre to attain a 15% return on assets rand / litre Internal rate of return Profitability of investing into biofuel production % Bottom line The overall bottom line cost implications of providing biofuel to the
Western Cape consumer through local production or inter-provincial imports rand
entities oppositely effecting one another, while vari-ables in a reinforcing loop influences each other to the same effect. This is best described through an example. In R1 an increase in population will lead to an increase in births, which again increases the population (thus reinforcing). An increase in popu-lation in B1, however, will lead to an increase in deaths, in turn decreasing the population and, con-sequently, balancing.
The CLD in Figure 2 shows the structure of the overall biofuel sector, by simplistically showing how fuel demand is driven by population growth. The mandatory blending policy, acting as an exogenous driver, creates a demand for biofuel, which in turn decreases fossil fuel demand and fossil fuel use. A demand for biofuel is shown to lead to biofuel pro-duction capacity, which results in biofuel produc-tion. Biofuel production is one of the main drivers and increases energy requirement, agricultural land use, water demand, by-product produced and employment, which ultimately leads back to an expansion of the biofuel production capacity (if bio-fuel production is financially lucrative). The CLD also indicates expected emissions being influenced by changes in fossil fuel use and the production of biofuel (through energy requirements and increased agricultural activities).
3.3 Dynamic modelling
The model-building phase resulted in the construc-tion of eight core sub-models. These models are described in detail by Jonker (2015), and include: biofuel production, agricultural yield, biofuel expen-diture, operational finances, profitability, alterna-tives to local production, employment, and emis-sions. Figure 3 shows an extract from the opera-tional finances sub-model, and is shown here to illustrate the key concept and working of system dynamics. From Figure 3 it can be seen how biofuel feedstock crop cost will increase annually with the biofuel feedstock cost increase, which is driven by a
percentage increase (feedstock increase rate) as well as the effects of demand and supply on market prices. The time horizon of the simulation is 2001 to 2040, where the 2001 census data provided accu-rate initial values for the model. This simulation time also provides sufficient historic simulation results in order to validate the model behaviour and outputs.
The biofuel model consists of eight sub-models, which are interconnected and represents the dynamics involved in biofuel production. The bio-fuel model further received input and information regarding agricultural activities, from a larger model (WeCaGEM), which describes the Western Cape’s transition to a green economy (Musango et al., 2015).
3.4 Scenario planning
In order for a model to be usefull and acurate, Senge (1980) mentioned that confidence in the model output has to be established. Numerous val-idation techniques are available to ensure that model structure and outputs are acurate, and the validation techniques as set out by Senge (1980) and Maani and Cavana (2007) were followed in this study. This included validating the model struc-ture with independent industry specialists, cross-ref-erencing model parameters with real-life expected values, and completing an extreme conditions and boundary adequacy test to ensure that the model responds logically. To further validate the model, dimensional consistency was checked and model behaviour was analysed against expected behavior and trends (Jonker, 2015).
Once the model was validated, a sensitivity analysis was done on variables that can be changed or determined endogenously. Electricty price was found to be one of the key endogenous variables influencing the success of local production (LP) sce-narios. Different project scenarios were thus gener-ated to look at the feasibility of generating electricity
on-site and incorporating electricity sales as part of the business case. As mentioned earlier, it was also necessary to look at the alternative solution for the Western Cape to form part of the mandatory blend-ing requirement, which will see the province pur-chasing biofuel from existing plants elsewhere in the country; in other words non-local production (NLP). The following six scenarios were identified and used as comparison to assesss the impacts and effects of complying with the mandatory blending requirement on a provincial level:
1. LP (boiler-coal): A local production scenario using a conventional coal-fired boiler to provide process heat.
2. LP (boiler-biomass): A local production scenario using a biomass-fired boiler for process heat. 3. LP (CHP-coal): A local production scenario
where a combined heat and power (CHP) unit using coal is installed to provide the production facility with process heat and electricty (where surplus electricty can be sold to the national grid).
4. LP (CHP-biomass): Biomass is used instead of coal to fuel the CHP mentioned in sceanrio 3. 5. NLP (coal): A non-local production sceanrio,
which evaluates the effects of buying biofuel in from a different province. To provide objective comparison of the emissions produced it is assumed that the said plant uses a coal fired boiler.
6. NLP (biomass): A non-local production scenario identical to scenario 5, but assuming the boiler is fuelled through the use of biomass.
4. Modelling outcomes
4.1 Biofuel production
The biofuel production sub-model simulated bioethanol and biodiesel production capacity and actual production volumes. The biofuel capacity model assumed that two facilities producing 160
million litres/annum of bioethanol and 35 million litres/annum of biodiesel will be completed by 2018 (GreenCape, 2015). Figure 4 shows biofuel produc-tion capacity and the actual biofuel producproduc-tion. The graph shows how actual production is less than pro-duction capacity and how, subsequently, produc-tion capacity decreases based on the reducproduc-tion in required production capacity due to biofuel feed-stock limiting the amount that can be produced.
Figure 5 shows the amount of cropland avail-able for biofuel crop production, based on the inputs received from the agricultural sub-model in WeCaGEM. The biofuel production model ass-sumed 25% of the current canola land to be allocat-ed to biodiesel production and a four-year crop rotation cycle for triticale cultivation to produce bioethanol. The effects of farmer uptake of triticale cultivation were incorporated and based on the sales price of triticale compared with B3-grade, or utility-grade, wheat, which is the lowest quality wheat grading, resulting in a substantially lower sell-ing price when compared with B1 grade. The farmer uptake is a function of the expected sales price of utility-grade wheat compared with the expected sales price of triticale, and it is assumed that, if the sales price of triticale were higher than than of utility grade wheat, the farmer would allo-cate (over time) the portion of land (based on his-torical fractions) yielding utility grade wheat to the cultivation of triticale. The model assumes that the sales price of triticale will initially be equal to that of utility grade wheat and therafter dynamically adjust based on supply and demand as illustrated by Figure 3. Due to triticale’s notable ability to grow in marginal soil, experts suggest that, in addition to crop rotation land available an additional 70 000 ha of uncultivated land is available for triticale cul-tivation in the Western Cape, which contributed to the production capacity (Amigun et al., 2011).
Figure 6 shows how the biofuel production under the current feedstock assumptions are not sufficient to meet the biodiesel demand of the province, based on fuel demand and mandatory blending requirements.
Figure 6: Expected biofuel shortage in the Western Cape.
4.1 Biofuel expenditure
Table 3 shows the expected annual operational expenditure, and it is evident that costs are expect-ed to increase drastically for both bioethanol and biodiesel production. It should be noted that the CHP scenarios (3 and 4) only have marginally high-er ophigh-erating costs than the boilhigh-er scenarios (1 and 2) by the year 2038, due to the dependence of operational cost on rising electricity costs.
The increase in operational expenditure is
driv-en by the increase in operational or input costs. Table 4 shows the various input costs where the low estimated cost of triticale and the high value of DDGS are some of the key factors driving the suc-cessful business case of bioethanol production. The simulation also highlights factors impeding the establishment of a biodiesel sector; these include the high cost of canola and the comparatively weak market value of biodiesel by-product. In addition, water, energy, labour and chemical costs are expect-ed to increase by 5–8% per annum, basexpect-ed on his-torical trends.
4.2 Profitability
Although it is commonly acknowledged that the biofuel industry brings advantages like a reduction in greenhouse gas emissions and employment cre-ation, attracting investment ultimately depends on the profitability of the industry. In order to attract investors, the national biofuel industrial strategy proposed a subsidy to be paid to producers to ensure a 15% return on assets (ROA) as defined by Equation 1 (DOE, 2014):
ROA = (1) The value of assets is initially estimated to be R900 million and R200 million for the respective bioethanol and biodiesel facilities, making use of grid-provided energy and a boiler for process heat (scenarios 1 and 2). Scenarios 3 and 4, including the CHP unit, are considerably more expensive because of the estimated installed capacity cost of ZAR 51 000 to ZAR58 000 per kW. This requires an investment of respectively R2 810 million and R2 120 million for bioethanol and biodiesel produc-tion facilities, including a 33 MW CHP. The value of
Table 3: Operational expenditure (R million).
Bioethanol Biodiesel 2018 2028 2038 2018 2028 2038 Feedstock cost 250 637 1083 60 128 235 Water cost 13 33 55 2 3 5 Chemical cost 25 86 194 8 27 62 Maintenance LP (boiler) 14 14 14 3 3 3 LP (CHP) 43 43 43 32 32 32 Labour 8 20 30 4 7 9 Energy cost LP (Boiler- Coal) 53 149 284 2 5 12 LP (Boiler - Biomass) 65 179 335 2 6 13 LP (CHP - Coal) 107 176 290 107 176 290 LP (CHP - Biomass) 142 234 385 142 234 385 OPEX LP (Boiler- Coal) 361 935 1658 75 161 304 LP (Boiler - Biomass) 373 966 1709 75 161 304 LP (CHP - Coal) 444 992 1693 209 361 611 LP (CHP - Biomass) 479 1050 1788 244 419 706
Earning before interest and taxes Total assets
assets is expected to decrease over the expected lifetime (20 years) of the facility. In addition to the subsidy required to achieve the guaranteed ROA, the model simulates the internal rate of return (IRR) for the various local production scenarios to assess the financial feasibility of biofuel production. Bioethanol shows a strong business case, requiring no subsidy to achieve the 15% ROA. Scenario 1 and 2 produced higher ROA values than scenario 3 and 4 due to the high capital cost of incorporating a CHP unit. The IRR values also indicated scenarios 1 and 2 to be the most lucrative, with IRR values of 25% and 23% respectively. For biodiesel produc-tion, none of the local production scenarios proved to be a feasible business case. Scenarios 1 and 2 required subsidies of between ZAR 4.30 and ZAR
1.70 per litre of biodiesel, while scenarios 3 and 4 require unrealistically high subsidies to reach the 15% ROA. The CHP scenarios again compared poorly with each other in the financial analysis, due to the substantial capital requirements. The IRR val-ues for all of the biodiesel scenarios were large neg-ative values over most of the project lifespan. 4.3 Bottom line cost
The bottom line cost refers to the overall cost of complying with the blending mandate and incorpo-rates scenarios 5 and 6 (non-local production), which will see the Western Cape buying biofuel from existing biofuel producers. The local and non-local production scenarios for bioethanol and biodiesel are shown in Figures 7 and 8. For
bio-Table 4: Operating costs.
Units 2018 2028 2038
Biofuel crop feedstock cost
Bioethanol – triticale Rand/ton 1 571 2 604 4 319
Biodiesel – canola Rand/ton 3 841 6 343 10 491
Technical and administrative wages Rand/person/hour 24 38 60
Management salary Rand/person/year 545 847 864 092 1 368 000
Water unit cost Rand/kilolitre 6.54 10.78 17.76
Combustible cost
Biomass Rand/ton 385 635 1 046
Coal Rand/ton 650 1 018 1 763
Electricity cost Rand/kWh 0.45 1 2.21
Chemical
Bioethanol – yeast Rand/kg 175.12 388.97 863.94
Bioethanol – enzyme Rand/kg 112.61 250.11 555.53
Biodiesel – methanol Rand/kg 4.66 10.35 22.99
Biodiesel – catalyst Rand/kg 8.54 18.97 43.14
By-product selling price
Bioethanol – DDGS Rand/ton 1 942 4 312 9 578
Biodiesel – canola oilcake Rand/ton 2 640 5 865 13 026
ethanol, the non-local production scenarios initially have lower associated costs, but will exceed the costs of all of the local production scenarios by 2025 because of the continued expenditure on pur-chasing and transporting bioethanol with very little income being generated by the non-local produc-tion scenarios. Figure 8 shows the bottom-line cost for biodiesel, where the cost of the local production scenarios making use of the CHP unit far exceeds the local production scenarios 1 and 2, as well as the non-local production scenarios. The non-local production bottom-line costs only marginally exceed scenarios 1 and 2 in 2036, meaning that for the majority of the estimated project lifespan it is expected to be cheaper to purchase biodiesel from existing production facilities, rather than producing biodiesel locally.
4.4 Employment
In evaluating employment created through a biofuel sector, only direct additional job creation was con-sidered. Employment in the biofuel production sec-tor can vary greatly, based on production method, scale and level of automation used. To simulate employment creation, historical data of employees required per area of cropland (33.64 employees/ hectare) farmed, and an employment ratio per litre of biofuel produced was incorporated into the model (Karaan, 2010; National biofuels study, 2006). It is estimated that the bioethanol plant will create roughly 2 080 jobs through expansion of the agricultural sector and an average of 55 additional jobs involved in operational duties. Biodiesel pro-duction creates relatively few jobs, with between 10 and 15 additional jobs expected, due to the scale of the operation and the use of existing canola crops (thus not increasing agricultural activity).
4.5 Emissions
The emissions produced by the various scenarios depend mainly on the process used and, as expect-ed, the scenarios making use of energy from coal showed substantially higher emissions. The bioethanol production facility utilising biomass for energy showed 63% less emissions than the coal-fired counterpart. Biodiesel production using biomass showed a 5% reduction in emissions when compared with a coal-fired facility due to the scale of the operation.
For bioethanol and biodiesel the non-local pro-duction scenarios showed higher emissions than local production, due to the additional emissions caused by transporting the biofuel from established facilities by road. Emissions were evaluated on the amount of carbon dioxide produced per litre of fuel burned; refer to Jonker (2015) for the detailed fac-tors and boundaries used to calculate the respective carbon footprints.
4.6 Scenarios summary and recommendations
All of the scenarios were evaluated according to the indicators identified (see Table 2). Recommend-ations to facilitate complying with the mandatory blending policy in the Western Cape are sum-marised below, and is discussed in detail in the work of Jonker (2015).
4.6.1 Bioethanol
The Local production scenario 2 (boiler-biomass) is recommended, as an IRR of 23% is achieved, while emissions are 63% lower than for the coal-fired option. To strengthen the business case of bioethanol production, the following factors should be addressed:
1. Improve feedstock availability through utilising uncultivated marginal land.
2. Reduce capital costs through alternative financ-ing options.
3. Incorporate bioethanol production as part of an invasive alien land-clearing scheme.
4.6.2 Biodiesel
It is recommended that biodiesel is not locally pro-duced (scenarios 5 and 6). The current operation scale and feedstock assumptions do not provide a compelling business case for local biodiesel produc-tion. Furthermore, the social and environmental advantages from biodiesel production does not pro-vide sufficient incentive to justify the financial and infrastructure burdens associated with a project of this scale and nature. Further work is recommended in the following areas to establish more feasible biodiesel production options:
1. Investigate different and combined feedstock options.
2. Consider systems and policies to encourage small-scale biodiesel production.
3. Re-evaluate the value stream of biodiesel. 4.7 Approach evaluation
System dynamics is a powerful modelling technique to holistically describe complex systems and is favoured for its versatile and robust nature. One of the disadvantages is, however, the extensive knowl-edge and data required on every part of the system in order to accurately isolate the root causes and drivers of the system. Some elements may be over-looked or simplified based on the dispositions of the stakeholders, although various stakeholders and experts were engaged to determine the system structure (see Jonker, 2015). An iterative and inter-active process is thus required to build a model that comprehensively captures the true nature of a com-plex system. Lack of available data and an estab-lished biofuel sector in South Africa (and particular-ly the Western Cape) necessitated a biofuel project evaluation model to be built. Such models generally
require the use of switches that inhibits the dynamic and feedback effects within the system. The largest drawback of the model is thus that it did not effec-tively incorporate feedback and dynamic linkages from the green economy model (WeCaGEM). Although it received inputs from WeCaGEM, the model did not indicate any emergent system behaviour. System dynamics was successfully used for the project evaluation of establishing a biofuel sector, but future work is recommended to incorpo-rate the biofuel model into a larger model in order to gain an in-depth understanding of the adverse effects of biofuel production as part of a green econ-omy transition.
5. Conclusions
This study investigated the literature available on policies in the biofuel sector in South Africa and evaluated different techniques suitable for the eval-uation of biofuel production as part of a larger com-plex system. A system dynamics model was devel-oped to describe and evaluate the various effects of complying with the mandatory biofuel blending requirement under different scenarios. After evalua-tion of the different scenarios, it was recommended that bioethanol is produced locally using biomass for process heat and biodiesel to be bought in from existing production plants if necessary. The useful-ness and applicability of system dynamics as a modelling technique for biofuel production as part of the green economy was investigated and it was concluded that it is extremely well suited for the rep-resentation of interrelated complex systems. System dynamics is not necessarily the best approach if one is purely interested in doing a project evaluation, but nonetheless provides valuable insight into the key drivers and intervention points for the evalua-tion of a system.
References
Amigun, B., Petrie, D. and Görgens, J. 2011. Economic risk assessment of advanced process technologies for bioethanol production in South Africa: Monte Carlo analysis. Renewable Energy 36(11): 3178–3186. Amigun, B., Petrie, D. and Görgens, J. 2012. Feedstock
and technology options for bioethanol production in South Africa: techno economic prefeasibility study. Energy &Fuels 26(9): 5887–5896.
Banos, R., Manzano-Agugliaro, F., Montoya, F., Gil, C., Alcayde, A. and Gomez, J. 2011. Optimization methods applied to renewable and sustainable ener-gy: A review. Renewable and Sustainable Energy Reviews 15(4): 1753 - 1766.
Bassi, A.M. 2014. Review of business and economics studies. Quarterly 2 (1): 88–97.
Brailsford, S., Churilov, L. and Dangerfield, B. 2014. Discrete-event simulation and system dynamics for management decision making. John Wiley and Sons.
Davis-Knight, H.R. and Weightman, R.M. 2008. The potential of triticale as a low input cereal for bioethanol production. Home-Grown Cereals Authority. Available online at: http://cereals.ahdb. org.uk/media/408618/pr434-final-project-report.pdf (accessed 14 January 2016.
Department of Energy (DOE) 2012. Government Gazette r. 671. Available online at: http://www.ener-gy.gov.za/files/policies/Mandatory%20Blending%20R egulations%2024%20August%202012.pdf (accessed 14 January 2016).
Department of Energy (DOE) 2014. Draft position paper on the South African biofuel regulatory frame-work. Government Gazette no. 37232. Available online at: http://www.gov.za/sites/www.gov.za/ files/37232gen24.pdf (accessed 14 January 2016). Department of Minerals and Energy 1998. White paper
on the energy policy of theRepublic of South Africa. Available online at: http://www.energy.gov.za/files/ esources/petroleum/wp_energy_policy_1998.pdf (accessed 14 January 2016).
Department of Minerals and Energy 2007. Biofuels industrial strategy of the Republic of South Africa. Available online at: http://www.energy.gov.za/files/ esources/renewables/biofuels_indus_strat.pdf(2).pdf (accessed 14 January 2016).
Epstein, R.J. 2014. A history of econometrics. Elsevier, North Holland.
Fore, S.R., Porter, P. and Lazarus, W. 2011. Net energy balance of small-scale on-farm biodiesel production from canola and soybean. Biomass and Bioenergy 35(5): 2234–2244.
GreenCape 2015. Biofuels: from viability to pilot project (re-assessment of the Western Capes biofuel produc-tion potential via a multi-criteria analysis). Cape Town. Available online at: http://greencape.co.za/ what-we-do/projects/biofuels/ (accessed 14 January 2016).
Helbing, D. and Balietti, S. 2013. How to do agent-based simulations in the future: From modelling social mechanisms to emergent phenomena and interactive systems design. In: Helbing D (ed). Social self-organization: Agent-based simulations and experiments to study emergent social behavior. Springer, Berlin: 25–70.
Jonker, W.D. 2015. Biofuel implications of a green econ-omy transition in the Western Cape province of South Africa: A system dynamics modelling approach to biofuel. Masters thesis, Stellenbosch University. Available online at: http://hdl.handle.net/ 10019.1/97865 (accessed 14 January 2016). Karaan, M. 2010. Can agriculture create a million jobs?
Paper presented at AgriSA Conference 2010, Muldersdrift, Gauteng, 8 October 2010.
Le Novère, N. 2015. Quantitative and logic modelling of molecular and gene networks. Nature Reviews Genetics 16(3): 146–58.
Loorbach, D. 2010. Transition management for sustain-able development: A prescriptive, complexity-based governance framework. Governance, 23(1): 161– 183.
Maani, K. and Cavana, R.Y. 2007. Systems thinking, system dynamics: Managing change and complexity. Prentice Hall.
Musango, J.K., Brent, A.C., van Niekerk, J.B.S., Jonker, W.D., Pienaar, A.P., York, T.A., Oosthuizen, J., Duminy, L. and de Kock, I.H. 2015. A system dynamics approach to understand the implications of a green economy transition in the Western Cape Province of South Africa. paper presented at the 33rdInternational Conference of the System
Dynamics Society, Boston, USA.
National Planning Commission 2012. National develop-ment plan 2030. Our future: Make it work. Pretoria. Available online at: http://www.gov.za/documents/ national-development-plan-2030-our-future-make-it-work (accessed 14 January 2016).
National Biofuels Study 2006. An investigation into the feasibility of establishing a biofuels industry in the Republic of South Africa. Available online at: http://www.mabelefuels.com/wp-content/uploads/ 2011/12/National-Biofuels-Study.pdf (accessed 7 December 2016).
Nolte, M. 2007. Commercial biodiesel production in South Africa: A preliminary economic feasibility study. Masters thesis, Stellenbosch University. Available online at: http://hdl.handle.net/10019.1/ 1797 (accessed 14 January 2016).
Pruyt, E. 2013. Small system dynamics models for big issues: Triple jump towards real-world complexity. Delft University of Technology. Available online at: http://simulation.tbm.tudelft.nl/smallSDmodels/Intro.h tml (accessed 14 January 2016).
Senge, P.M. 1980. Tests for building confidence in sys-tem dynamics models. Syssys-tem dynamics TIMS Studies in Management Sciences 14: 209-228. South African Petroleum Industry Association (2015
June). Industry overview – South African fuel indus-try. Available online at: http://www.sapia.co.za/indus-try-overview/fuel-industry.html (accessed 18 January 2017).
Snijders, T.A., van de Bunt, G.G. and Steglich, C.E. 2010. Introduction to stochastic actor-based models for network dynamics. Social Networks 32(1): 44– 60.
United Nations Environment Programme (UNEP). 2010. Driving a green economy through public finance and fiscal policy reform. Paris. Available online at: http://www.unep.org/greeneconomy/Portals/88/docu-ments/ger/GER_Working_Paper_Public_Finance.pdf (accessed 14 January 2016).
Western Cape Government (2013). Green is smart. Cape Town. Available online at: https://www.west-erncape.gov.za/110green/tags/green-smart (accessed 14 January 2016).
